Neurogenesis and epilepsy in the developing brain
Address correspondence to Brenda E. Porter, The Children's Hospital of Philadelphia, University of Pennsylvania, 502c Abramson Research Building, 3516 Civic Center Boulevard, Philadelphia, PA 19104, U.S.A. E-mail: email@example.com
Multiple studies have highlighted how seizures induce different molecular, cellular, and physiologic consequences in an immature brain as compared to a mature brain. In keeping with these studies, seizures early in life alter dentate granule cell birth in different, and even opposing, fashion to adult seizure models (seeTable 1). During the first week of rodent postnatal life, seizures decrease cell birth in the postictal period, but do not alter the maturation of newborn cells. Seizures during the second week of life have varied effects on dentate granule cell birth, either causing no change or increasing birth, and may promote a mild increase in neuronal survival. During the third and fourth weeks of life, seizures begin to increase cell birth similar to that seen in adult seizure models. Interestingly, animals that experienced seizure during the first month of life have an increase in cell birth during adulthood, opposite to the reported decrease in chronic animals experiencing a prolonged seizure as an adult. Children have more ongoing cell birth in the dentate gyrus than adults, and markers of cell division are further increased in children with refractory temporal lobe epilepsy. There are clear age-dependent differences in how seizures alter cell birth in the dentate gyrus both acutely and chronically. Future studies need to focus on how these changes in neurogenesis influence dentate gyrus function and what they imply for epileptogenesis and learning and memory impairments, so commonly found in children with temporal lobe epilepsy.
Table 1. Early seizures alter cell birth in rodent dentate gyrus
|1st week of life||Flurothyl||Increased ictala and decreased postictalb||ND|
|2nd week of life||Pilocarpine||Increasede||ND|
|Hyperthermia||No changef,g||Slight increase in newborn cell survivalg (female only)|
|3rd –4th week of life||Pilocarpine||Increasede,h||Increased in those animals that develop spontaneous seizuresj|
|Kainate||Increasedi(similar to adults)||ND|
The robust alteration in neurogenesis following an episode of status epilepticus is an interesting phenomenon and begs the question of how this change might contribute to epilepsy as well at temporal lobe dysfunction. In early-life seizure animal models and in children that experience early-life seizures, there is a lower seizure threshold and an increased risk for developing epilepsy and cognitive dysfunction later in life. Here, we begin to address how changes in regulation of cell birth in the dentate gyrus following seizures during development might contribute to these problems. We start by describing how seizures alter dentate gyrus neurogenesis during rodent development and then turn to what is known about neurogenesis in children with epilepsy.
The majority of cortical neurogenesis occurs early during prenatal development: 16 weeks gestation in humans and by 15 days postfertilization in rodents. The hippocampus and cerebellum, however, are late-developing neuronal structures, completing formation after 34 weeks of gestation in humans and during the first 2 weeks of postnatal life in the rat (Bayer, 1980; Arnold & Trojanowski, 1996; Hauser et al., 2003). Therefore, if seizures have an effect on brain development, it is most likely to be in these late-developing neuronal structures. Hence, in premature infants and immature rodent epilepsy models, seizures early in life might influence neurogenesis that is required for the formation of the hippocampus and cerebellum. Here, we discuss the effects of seizures early in life on cell birth in the dentate gyrus, both at acute and chronic time points, and what is known about how the seizures alter maturation of the newborn cells.
Postnatal Week 1 in Rat
The work by Wasterlain et al. in the 1970s suggested that seizures within the first week of life decrease the rate of cell birth in the brain, though they did not identify the primary cell type affected (Wasterlain & Plum, 1973; Wasterlain, 1975). More recently, focusing on the effect of multiple brief seizures during the first postnatal week, the laboratory of Dr. Greg Holmes showed that seizures both promote and suppress cell birth in the dentate gyrus depending on when cell birth is measured in relationship to the seizure (Holmes et al., 1998; McCabe et al., 2001). The timing of the thymidine analog injection, bromo-deoxyuridine (BrdU), determined whether the seizure appeared to increase or decrease cell birth in the dentate gyrus. If BrdU was given just prior to a brief flurothyl-induced seizure, there was a small increase in BrdU labeling, suggesting that, during the seizure, BrdU incorporation and cell birth in the dentate gyrus are increased. In contrast, if the BrdU was injected following multiple flurothyl-induced seizures, there was a decrease in BrdU labeling. Thus, cell birth is suppressed in the postictal period following multiple brief seizures during the first week of life. They confirmed that the postictal decrease was age-related and found an increase in dentate granule cell birth following multiple flurothyl-induced seizures in mature animals. In immature animals, the decrease in BrdU labeling persisted for 2 weeks, and the majority of labeled cells in both the control and flurothyl-treated rats eventually expressed mature neuronal markers, suggesting that the seizure did not alter differentiation of the newborn cells. Consistent with the findings of the Holmes laboratory, Xiu-Yu et al. (2007) used repetitive doses of pilocarpine to induce seizures during the first week of life and found an approximately 40% reduction in BrdU labeling for the weeks following seizure induction. Liu et al. injected kainate at several ages from P6 to P13 to cause multiple seizures and found a decrease in cell birth in the postictal period (Liu et al., 2003). Taken together, seizures during the first week of life, in contrast to seizures in adults, repress cell birth during the postictal period in the dentate gyrus, with suppression persisting for several weeks after the seizure.
The mechanism for postictal suppression of neurogenesis following seizures in the first week of life is not understood. One possible explanation is that seizures interact differently with the initial wave of neurogenesis important for the formation of the dentate gyrus versus the ongoing replacement of neurons found at later ages. Alternatively, seizures early in life produce distinct physiologic and molecular changes that are age-specific and might be responsible for the unique decrease in cell birth during the first week of life.
Interestingly, decreased cell birth in the dentate gyrus lasts for 2 weeks after the pilocarpine-induced seizures and is then followed by an increase in cell birth at chronic time points. These data contrast with decreased neurogenesis at chronic time points in adult seizure models (Hattiangady et al., 2004; Kralic et al., 2005). Further studies are needed to determine if the chronic increase in cell birth following the seizures during the first week of life is found in other seizure models such as that of flurothyl. These findings imply that seizures during the first week of life have very different short- and long-term influences on cell birth and neurogenesis compared to the older age groups.
Postnatal Week 2 in Rat
During the second week of life, the chemoconvulsants kainic acid and lithium-pilocarpine and the hyperthermia models of febrile seizures have been used to induce seizures and measure cell birth in the dentate gyrus. The findings across models are not consistent; for example, kainate and pilocarpine seizures increase neurogenesis during the postictal period (Sankar et al., 2000; Bender et al., 2003). Conversely, in the P10 hypertherthermia-induced seizure model, there was no immediate postictal change in cell birth in the week following the seizure (Bender et al., 2003; Lemmens et al., 2005). However, if the hyperthermia-treated animals were labeled with BrdU during the week following the seizure and sacrificed in adulthood, there was a modest increase in labeled cells in the female, but not in the male, littermates (Lemmens et al., 2005). This suggests a mild increase in long-term dentate granule neuron survival following P10 hyperthermia-induced seizures in female rats. It is not known if there are differences in the long-term survival and/or maturation of newborn cells following kainate- and pilocarpine-induced seizures at this age.
The method of seizure induction appears to be important in determining if and how seizures influence cell birth. Why there is a strong difference in how each method influences cell birth in the dentate gyrus is not understood, but may relate to the mechanism of seizure induction. The difference in neurogenesis among P10 kainate-, pilocarpine-, and hyperthermia-induced seizures allows for speculation about the role neurogenesis plays in epileptogenesis. Animals that experienced lithium-pilocarpine-provoked seizures at P10 do not develop spontaneous seizures, though they do have lowered seizure thresholds for subsequent chemoconvulsant challenges (Dube et al., 2001a; Zhang et al., 2004a, 2004b). In contrast, hyperthermia-induced seizures at P10 cause a subset of animals, approximately one-third, to develop brief spontaneous seizures in adulthood (Dube et al., 2006). The lack of altered neurogenesis following hyperthermia-induced seizures suggests that altered cell birth does not contribute to epileptogenesis in these animals. Further studies are needed to determine if the increase in neurogenesis in the pilocarpine and kainate models at this age contributes to a lowered seizure threshold or learning and memory differences in adulthood (Holmes, 2005).
Postnatal Week 3–4 in Rat
At what age do seizures in rodents induce changes in cell birth and differentiation that is similar to adults? Kainate-induced seizures at 1 month and at 3 months of age induce similar increases in cell birth (Gray et al., 2002). During the third postnatal week, pilocarpine chemoconvulsants appear to induce an increase in cell birth that may be slightly more robust than P10 and closer to that of the adult; however, dosing and methodological differences make it difficult to be certain. There is a several-fold increase in cell birth following lithium-pilocarpine-induced seizures at P20 (Sankar et al., 2000; Porter et al., 2004). Akman et al. found an immediate decrease in neurogenesis following lithium-pilocarpine-induced seizures at P20. The difference may be model-specific as Akman et al. used diazepam to stop the seizure at 2 h, and γ-aminobutyric acid (GABA) agonists have been reported to decrease neurogenesis (Akman et al., 2004; Tozuka et al., 2005). Use of GABA agonists in the adult pilocarpine model may explain the more immediate increase in cell birth in the P10 and P20 lithium-pilocarpine model that do not use benzodiazepines or barbiturates to stop status epilepticus (Parent et al., 1997; Sankar et al., 2000).
In addition to the increased cell birth following lithium-pilocarpine-induced status epilepticus at P20, there is also an increase in dying immature and mature dentate granule neurons, resulting in approximately three-fold more immature dentate granule neurons in the weeks following the seizure (Porter et al., 2004). How seizures increase the birth of dentate gyrus cells following a seizure is not well understood and requires further study. For instance, it is not known if dentate granule neurons birth and death are synchronized. A feedback mechanism seems likely as the size of the dentate gyrus does not dramatically change throughout life.
The immature neurons born following the P20 pilocarpine-induced seizure express a variety of neurotransmitters receptors including AMPA and kainite receptors that are relatively unaffected by prolonged seizures (Porter et al., 2006). In contrast, the mature neurons likely present during the episode of status epilepticus undergo a variety of changes in their AMPA and kainate receptor expression. One hypothesis from these data is that the immature cells offer a pool of cells with a seizure-naive phenotype for potential manipulation to suppress epileptogenesis or improve learning and memory.
Following lithium-pilocarpine-induced status epilepticus at postnatal day 20, approximately 50% of the animals go on to develop spontaneous seizures (Dube et al., 2001a, 2001b; Raol et al., 2003). The adult animals that eventually develop spontaneous seizures have an approximately two-fold increase in cell birth in the dentate gyrus as compared to those without spontaneous seizures or control animals (Cha et al., 2004). The chronic spontaneous seizure animals did not experience seizures in the period immediately preceding the BrdU injections, but it was not certain if they remained seizure-free for the duration of the experiment following the BrdU injections. This chronic increase in cell birth in animals developing spontaneous seizures following status epilepticus at P20 is opposite to the decrease in cell birth in chronic adult epilepsy models (Hattiangady et al., 2004; Kralic et al., 2005). Memory impairment has been found in adult animals following lithium-pilocarpine-induced status epilepticus at P20, but it is not known if the increase in cell birth is related to the poor memory performance (Liu et al., 1994).
Changes in Cell Birth and Neurogenesis in Children with Epilepsy
A number of studies have tried to assess cell birth and neurogenesis in epilepsy surgery specimens from children with medically refractory epilepsy and tumor-related epilepsy. Because BrdU cannot be used as a marker of cell division in human surgical and postmortem tissue, researchers have had to use biomarkers of stem cell number and cell division such as the proteins nestin and Ki-67, and the immature dentate granule neuron marker PSA-NCAM. Because nestin, Ki-67, and PSA-NCAM are only biomarkers, age or seizures might alter antigen expression without an actual change in stem cell number, cell division, or immature neurons. In autopsy control specimens and epilepsy surgical specimens, stem cell and dividing cell numbers appear to be greatest early in life and fall with age (Blumcke et al., 2001; Fahrner et al., 2007). Epilepsy with evidence of hippocampal pathology such as Ammon's horn sclerosis in infants and young children is associated with increased numbers of stem cell and dividing cells compared to age-matched controls (Blumcke et al., 2001; Takei et al., 2007). The increase in stem cell and dividing cells in pediatric patients with epilepsy and hippocampal pathology is similar to most (Crespel et al., 2005; Thom et al., 2005), but not all (Fahrner et al., 2007), adult epilepsy histopathologic case series. In contrast to the increase in markers of stem cell and cell division in children with epilepsy, there was a decrease in PSA-NCAM staining intensity, a marker of immature dentate granule neurons in pediatric epilepsy surgical specimens (Mathern et al., 2002). If confirmed with other markers of immature dentate granule neurons, such as doublecortin, the presence of increased total and dividing stem cells in the dentate gyrus of pediatric temporal lobe epilepsy patients suggests a loss or diversion of cells on the path to dentate granule neuronal maturation.
Seizures at all stages of development alter cell birth and neurogenesis in the rodent dentate gyrus (see Table 1). The age at the time of the seizure dramatically alters how the seizure influences cell birth and neurogenesis both acutely and chronically. Multiple seizures types during postnatal week 1 of life cause a prolonged postictal decrease in cell birth and neurogenesis. During the second postnatal week of life, some seizure types do not alter cell birth, while others cause a postictal increase. Seizures during the third and fourth weeks of life cause an increase in cell birth and neurogenesis. Animals exposed to a variety of seizure types during the first month of life have a mild increase in cell birth when evaluated in adulthood. A better understanding of these age-dependent differences in seizure-induced alteration in neurogenesis will help in determining the role of neurogenesis in epileptogenesis and learning and memory impairments in epilepsy.
Over the past few years, neuroscientists have made progress in describing the phenomenon of seizure-induced alterations in dentate gyrus neurogenesis during development. In the future, researchers will need to go beyond these descriptive studies and begin to determine if changes in neurogenesis are important for epileptogenesis and temporal lobe dysfunction. The mechanism by which seizures alter neurogenesis will need to be determined, as well as the large age-dependent and model-specific differences in how seizures influence neurogenesis. Understanding the mechanism by which seizures alter neurogenesis will have far-reaching implications not only for epilepsy but also in the fields of depression and memory impairment.
Conflict of interest: I confirm that I have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The author declares no conflict of interest.